Xplor3.843 map.

Gerard 'CD' Kleywegt GERARD at XRAY.BMC.UU.SE
Wed Nov 20 14:09:15 EST 1996


sigma levels of difference maps are *MEANINGLESS*
consider this:
at the beginning of your refinement, your model is
incomplete and contains errors -> lot of real signal
in the difference map and even a "1.5 sigma" peak may
be highly significant
at the end of refinement, by definition your difference
map is essentially flat, and any variation is random
noise; nevertheless there may still be "3 sgima" peaks
of course, only they don't mean anything
so, contour your difference map at some Electron/A3
level rather than at "N sigma" levels

of course your ligand could still be partly occupied;
you could see how the maps (and temp. factors) look
if you put it at half occupancy (if you have very
high resolution data you can also try to refine
occupancies)

--gerard



More information about the X-plor mailing list